Caleb Welton <cwelton@greenplum.com> writes:
> On 12/1/09 7:38 PM, "Tom Lane" <tgl@sss.pgh.pa.us> wrote:
>> Under what circumstances would users (or anyone at all) be putting data into an int2vector?
> What exactly is your objection to having the int2arrayin parser handle its input conversion reasonably?
I'm trying to gauge what the actual use-case is for having a slightly
nicer error behavior. The proposed patch adds another translatable
error string, which is no skin off my own nose but does create ongoing
work for our translation team. And presumably, if we're going to fix
this, we ought to fix the about-equally-stupid parsing logic in oidvectorin.
While we're at it, should we trouble to detect overflow in int2vectorin?
You could spend quite a bit of time and code making these functions more
bulletproof, but I'm not convinced it's worth any work.
regards, tom lane